CURUMIM: A Serendipitous Recommender System for Tourism Based on Human Curiosity

Alan Menk, Laura Sebastia, Rebeca Ferreira

Abstract: In recent years, an evolution of recommendation systems has been observed. New advisory systems have been gaining space, using new tools, algorithms and recommending techniques, not only to increase accuracy, but also to positively surprise the user, that is, to provide serendipitous recommendations. Following this trend, we describe CURUMIM, an online tourism recommender system, able to generate serendipitous recommendations of places around the world. In summary, from data available on social networks, it predicts the degree of curiosity of a user, which is then used, along with the user history and her level of education, to select the most appropriate recommendations. We have performed an experiment with real users who reported positive levels of satisfaction with the recommendations in terms of accuracy, serendipity and novelty.